executive summary why we did this review · executive summary tallahassee memorial healthcare,...

EXECUTIVE SUMMARY

Tallahassee Memorial HealthCare, Inc., submitted quality measure data to the Centers for Medicare & Medicaid Services that was not always complete.

WHY WE DID THIS REVIEW

The HospitalCompare.:Qhs.gov Web site (HospitalCompare) provides the public with information about the quality of care at over 4,000 Medicare-certified hospitals across the country. Consumers can use HospitalCompare to choose a health care provider by comparing the quality of care provided by various hospitals. Between July 1, 2009, and June 30, 2010, the public viewed HospitalCompare more than 13 million times. Therefore, it is important that the hospital quality measure data, through which HospitalCompare makes its comparisons, be complete and accurate.

The objective of our audit was to determine whether Tallahassee Memorial HealthCare, Inc. (the Hospital), submitted complete quality measure data for the period July 1, 2009, through June 30, 2010.

BACKGROUND

Section 501(b) of the Medicare Prescription Drug, Improvement, and Modernization Act of2003 required the submission of hospital quality measures. The Deficit Reduction Act of 2005 required the Department of Health and Human Services to establish procedures for making data submitted for the hospital reporting program available to the public. In response to this requirement, the Centers for Medicare & Medicaid Services (CMS) established HospitalCompare.

CMS bases the quality measures used to compare providers on HospitalCompare on scientific evidence that reflects guidelines, standards of care, or practice parameters. A quality measure converts medical data from patient records into a rate or percentage that allows both hospitals and consumers to evaluate a specific hospital's performance. These quality measures reflect the quality of care that patients receive during inpatient visits to hospitals. Hospitals may review their data before it is made available to the public. It takes 9 months or more before a new quarter of quality of care data affects the quality measures reported on HospitalCompare.

The Specifications Manual for National Hospital Inpatient Quality Measures (the Manual) contains the CMS requirements for the hospital reporting program. The Manual contains detailed instructions, diagnosis code tables, and calculation algorithms that hospitals use in collecting, abstracting, and submitting data to the "Clinical Data Warehouse" (Warehouse) for each quality measure. The Warehouse is a data repository that contains data uploaded from hospitals across the nation. The intent of this repository is to reduce the burden of duplicate data collection and reporting, provide comparative data on a State and national level, and prepare hospitals for national initiatives that are underway for quality information reporting.

Tallahassee Memorial HealthCare, Inc. , Quality Measure Data (A-04-11-08002)

Tallahassee Memorial HealthCare, Inc., is a 770-bed acute care hospital located in Tallahassee, Florida. The Hospital contracts with a vendor for software and data submission services for the hospital reporting program. From July 1, 2009, through June 30,2010, the Hospital submitted to the Warehouse data on 590 heart failure cases, 561 acute myocardial infarction cases, and 425 pneumoma cases.

WHAT WE FOUND

The Hospital submitted quality measure data that was not always complete. Specifically, the Hospital submitted incomplete data for pneumonia cases. These errors occurred because of an undetected coding error in the Hospital's database. In addition, the Hospital did not have sufficient written policies and procedures necessary for effective controls over quality measure data submissions. As a result, the Hospital's information that was available to the public on HospitalCompare was not always reliable.

WHAT WE RECOMMEND

We recommend that the Hospital improve controls over quality measure data by establishing:

• controls to ensure that data submitted to the Warehouse are complete and in accordance with Manual specifications and

• written policies and procedures, separate from those used by its vendor, that address the completeness of data submitted to the Warehouse.

TALLAHASSEE MEMORIAL HEALTHCARE COMMENTS

In written comments on our draft report, the Hospital agreed with the findings above and has taken steps to implement the recommendations. However, the Hospital did not agree with the reasons that we initially cited for the incomplete data submissions, and also did not agree with a finding that we included in our draft report regarding inaccurate quality measure submissions.

OFFICE OF INSPECTOR GENERAL RESPONSE

Based on the Hospital's written comments and further discussions, we modified the reasons that we cited for the incomplete data submissions and removed a finding from the draft report involving inaccurate quality measure data. The remaining recommendations are unchanged.

Tallahassee Memorial HealthCare, Inc. , Quality Measure Data (A-04-11-08002) ii

TABLE OF CONTENTS

INTRODUCTION ........................................................................................................................ !

Why We Did This Review ................... ............................................................................. 1

Objective ............ .............................. .......... ....................... ................................................ !

Background............................................... ........................................................................ I

Hospital Inpatient Quality Reporting Program ...................... ............................... 1

Prior Work by the Government Accountability Office ......... ............................... 2

Tallahassee Memorial HealthCare, Inc................................. ............................... 2

How We Conducted This Review .................................................................................... 3

FINDINGS .. .... .. ... ..... .. .. ..... .. .. .... .. .. ...... .... .... .... .......... .... ....... .......... ... ........ .. .. ............. .... .. ....... .... 3

Incomplete Quality Measure Data Submissions .............................................................. 3

Inadequate Policies and Procedures.................................................................................. 4

RECOMMENDATIONS .................................................................................... .......................... 4

TALLAHASSEE MEMORIAL HEALTHCARE COMMENTS ................................................4

OFFICE OF INSPECTOR GENERAL RESPONSE ............................ ....................................... 5

APPENDIXES

A: Federal Requirements ......................... .... .......... ........................... ............................... 6

B: Process of Care Quality Measures ..... .............................. .......... ................................. 9

C: Audit Scope and Methodology ....... .... ........ ................................. ...... .... ...... .... ..... ... 11

D: Tallahassee Memorial HealthCare Comments .......................... ... ..... ....................... 13

Tallahassee Memorial HealthCare, Inc. , Quality Measure Data (A-04-11-08002) iii

INTRODUCTION

WHY WE DID THIS REVIEW

The HospitalCompare.hhs. gov Web site (HospitalCompare) provides the public with information about the quality of care at over 4,000 Medicare-certified hospitals across the country. Consumers can use HospitalCompare to choose health care providers by comparing the quality of care provided by various hospitals. Between July 1, 2009, and June 30,2010, the public viewed HospitalCompare more than 13 million times. Therefore, it is important that the hospital quality measure data, through which HospitalCompare makes its comparisons, be complete and accurate.

OBJECTIVE

Our objective was to determine whether Tallahassee Memorial HealthCare, Inc. (the Hospital), submitted complete quality measure data for the period July 1, 2009, through June 30, 2010.

BACKGROUND

Hospital Inpatient Quality Reporting Program

Section 501(b) of the Medicare Prescription Drug, Improvement, and Modernization Act of2003 required the submission ofhospital quality data. The Deficit Reduction Act of 2005 required the Department of Health and Human Services to establish procedures for making data submitted for the hospital reporting program available to the public. 1 In response to this requirement, the Centers for Medicare & Medicaid Services (CMS) established Hospital Compare.

CMS bases the quality measures used to compare providers on HospitalCompare on scientific evidence that reflects guidelines, standards of care, or practice parameters. A quality measure converts medical data from patient records into a rate or percentage that allows both hospitals and consumers to evaluate a specific hospital's performance. These quality measures reflect the quality of care that patients receive during inpatient visits to hospitals.

Hospitals may review their data before it is made available to the public. It takes 9 months or more before a new quarter of quality of care data affects the quality measures reported on HospitalCompare. The quality measure data on HospitalCompare consist of the average of 4 rolling quarters (1 year's worth) of inpatient hospital data, as reported by the hospitals to CMS. Therefore, every quarter the oldest quarter data rolls out and the newest quarter data rolls in before CMS recalculates each measure.

The Specifications Manual for National Hospital Inpatient Quality Measures (the Manual) contains the CMS requirements for the hospital reporting program. The Manual contains detailed instructions, diagnosis code tables, and calculation algorithms that hospitals use in

1 § 5001 (a)(3)(viii)(VII) of the Deficit Reduction Act of2005 (P.L. 109-171 ).

Tallahassee Memorial HealthCare, Inc. , Quality Measure Data (A-04-/ 1-08002) 1

2 collecting, abstracting, and submitting data to the "Clinical Data Warehouse" (Warehouse) for each quality measure. The Manual also includes a "Data Dictionary" that describes the data elements required to report various measurements. CMS maintains the Warehouse and uses the submitted data to calculate the performance information reported on HospitalCompare. An example of specifications for one Pneumonia quality measure, Pneumonia Patients Assessed and Given Pneumonia Vaccination, is included in Appendix A.

HospitalCompare reports quality measure data, among other things, on heart failure, acute myocardial infarction, and pneumonia. These 3 diagnoses are among the 10 most common for Medicare inpatient care and, therefore, have a significant effect on Medicare program costs. Determination whether a particular case is appropriate for measurement is based on, among other factors, diagnosis codes at the time of discharge from the hospital. A list of the quality measures for heart failure, acute myocardial infarction, and pneumonia is detailed at Appendix B.3

Prior Work by the Government Accountability Office

The Government Accountability Office (GAO) has previously testified before the Committee on Finance, United States Senate, that CMS had no ongoing process for ensuring completeness of quality data.4 In an earlier report, GAO stated that, for hospital quality data to be useful to patients and other users, it needs to be reliable. 5 If a hospital did not collect or abstract the data from the patient's medical records accurately, the data would not be reliable. Similarly, if a hospital submitted accurate data, but those data were incomplete because the hospital leaves out eligible cases, the data would not be reliable.

Tallahassee Memorial HealthCare, Inc.

Tallahassee Memorial HealthCare, Inc., is a 770-bed acute care hospital located in Tallahassee, Florida. The Hospital contracts with a vendor for software and data submission services for the hospital reporting program.6 The Hospital uploads case discharge details to the vendor's software. The software filters the patient information based on the Manual requirements and provides a listing of cases to the Hospital for data abstraction. The Hospital has designated staff responsible for data transmission and data abstraction processes. After the data have been

2 The Clinical Data Warehouse is a data repository that contains data uploaded from hospitals across the nation. The intent of this repository is to reduce the burden of duplicate data collection and reporting, provide comparative data on a state and national level, and prepare hospitals for national initiatives that are underway for quality information reporting.

3 We limited the specific set of quality measures included in the scope of this audit to the process of care quality measures defined and listed in Appendix B.

4 GA0-08-555T, March 2008, Testimony Before the Committee on Finance, United States Senate, Hospital Quality Data, Issues and Challenges Related to How Hospitals Submit Data and How CMS Ensures Data Reliability.

5 GA0-06-54, January 2006, Hospital Quality Data, CMS Needs More Rigorous Methods to Ensure Reliability of Publicly Released Data.

6 The vendor is Thomson Reuters.

Tallahassee Memorial HealthCare, Inc. , Quality Measure Data (A-04-11-08002) 2

abstracted, the Hospital's vendor submits the data to the Warehouse. For the 4 quarterly periods

of July 1, 2009, through June 30, 2010, the Hospital submitted to the Warehouse data on 590

heart failure cases, 561 acute myocardial infarction cases, and 425 pneumonia cases.

Appendix B shows the quality measures reported on HospitalCompare.

HOW WE CONDUCTED THIS REVIEW

Our audit covered the Hospital's data submitted and used for quality measures associated with heart failure, acute myocardial infarction, and pneumonia during the period July 1, 2009, through June 30, 2010.

We conducted this performance audit in accordance with generally accepted government auditing standards. Those standards require that we plan and perform the audit to obtain sufficient, appropriate evidence to provide a reasonable basis for our findings and conclusions based on our audit objectives. We believe that the evidence obtained provides a reasonable basis for our findings and conclusions based on our audit objectives.

Appendix C contains the details of our audit scope and methodology.

FINDINGS

The Hospital submitted quality measure data that was not always complete. Specifically, the Hospital submitted incomplete data for pneumonia cases. These errors occurred because of an undetected coding error in the Hospital's database. In addition, the Hospital did not have sufficient written policies and procedures necessary for effective controls over quality measure data submissions. As a result, the Hospital's information that was available to the public on HospitalCompare was not always reliable.

INCOMPLETE QUALITY MEASURE DATA SUBMISSIONS

The Hospital's data submissions were incomplete for pneumonia cases. Specifically, the Hospital's controls did not identify 14 cases that had a Principle Diagnosis Code of Septicemia or Respiratory Failure that should have been submitted to the Warehouse with other pneumonia

7quality measure cases. The Hospital determined that these cases had been uploaded to its vendor software, but the Hospital's data processing system resequenced the diagnosis codes associated with the cases, causing the vendor ' s software not to recognize these 14 cases for submission to the Warehouse.

After we notified the Hospital of the potential case omissions in 2009 and 2010 data, the Hospital said that it identified similar errors in 2011 data that it could not correct because the cutoff dates for abstracting and uploading the data to the warehouse had passed. The Hospital identified additional case omissions in 2012 data and, because the cutoff dates had not yet passed, corrected the omitted cases identified in its 2012 data.

7 The omitted pneumonia cases that we identified may not represent all omitted pneumonia cases, but we found no omissions for heart failure and acute myocardial infarction cases in the Hospital ' s submissions to the Warehouse.

Tallahassee Memorial HealthCare, Inc. , Quality Measur e Data (A-04-11-08002) 3

This error occurred because of an undetected coding error in the Hospital's database.

As a result, the Hospital could not ensure that the information made available to the public on HospitalCompare was complete with respect to the pneumonia quality measures.

INADEQUATE POLICIES AND PROCEDURES

Although the Hospital had designated staff for data processing and abstraction functions , it did not have adequate written policies and procedures to control the completeness of data submitted to the Warehouse. Specifically, Hospital staff prepared informal procedures to extract case information from the Hospital's database of quarterly discharges for submission to the vendor. However, these procedures were insufficient to ensure complete data submissions. The Hospital had inadequate written policies and procedures in part because it relied on vendor software for quality data submissions under the hospital reporting program and was not aware that additional written policies and procedures were necessary for effective controls.

As a result of the lack of clear, written policies and procedures, the Hospital could not ensure that information submitted to the Warehouse was always complete. We could not determine the extent of incomplete data on the Hospital's calculated percentages reported on HospitalCompare because we could not be sure that we identified all data omissions for pneumonia, heart failure, and acute myocardial infarction.

RECOMMENDATIONS

We recommend that the Hospital improve controls over quality measure data by establishing:

• controls to ensure that data submitted to the Warehouse are complete and in accordance with Manual specifications and

• written policies and procedures, separate from those used by its vendor, that address the completeness of data submitted to the Warehouse.

TALLAHASSEE MEMORlAL HEALTHCARE COMMENTS

In written comments on our draft report, the Hospital agreed with the findings above and has taken steps to implement the recommendations. However, the Hospital did not agree with the reason that we initially cited for the incomplete data submissions and also did not agree with a finding that we included in our draft report regarding inaccurate quality measure submissions.

Regarding incomplete quality measure data submissions, the Hospital indicated that it has added steps to the data collection process to prevent omission errors from reoccurring. The Hospital corrected the software code error so diagnosis codes are no longer re-sequenced prior to processing. The Hospital implemented additional queries for reportable cases based on diagnosis code to en sure the completeness of the data reported.

Tallahassee Memorial HealthCare, Inc., Quality Measure Data (A-04-11-08002) 4

Regarding inadequate policies and procedures, the Hospital is in the process of creating a written policy and procedure to address the submission of data from the Hospital to the Warehouse including the use of its vendor. This policy and procedure will encompass the steps necessary to completely gather data from the various hospital information systems, verify the accuracy of the data gathered, and ensure transmission of the data is accomplished.

The Hospital's written comments on the draft report are included in their entirety as Appendix D.

OFFICE OF INSPECTOR GENERAL RESPONSE

Based on the Hospital's written comments and further discussions, we modified the reasons that we cited for the incomplete data submissions and removed a finding from the draft report involving inaccurate quality measure data. The remaining recommendations are unchanged.


APPENDIX A: FEDERAL REQUIREMENTS

Hospital Reporting Program

Section 50l(b) of the Medicare Prescription Drug, Improvement, and Modernization Act of2003 required the submission of hospital quality data.

Deficit Reduction Act of2005, section 5001(a): amended the Social Security Act section 1886(b)(3)(B)(viii) by adding ... new clauses:

(II) Each subsection (d) hospiial shall submit data on measures selected under this clause to the Secretary in a form and manner, and at a time, specified by the Secretary for purposes of this clause ....

(VII) The Secretary shall establish procedures for making data submitted under this clause available to the public. Such procedures shall ensure that a hospital has the opportunity to review the data that are to be made public with respect to the hospital prior to such data being made public. The Secretary shall report quality measures of process, structure, outcome, patients' perspectives on care, efficiency, and costs of care that relate to services furnished in inpatient settings in hospitals on the Internet Web site of the Centers for Medicare & Medicaid Services.

The Manual:

"Alphabetical Data Dictionary," version 3.1, page 1-204, indicates that the Data Element "hospital patient identifier" is collected for all records and is defined as the number used by the hospital to identify this patient's stay. The number provided will be used to identify the patient in communications with the hospital, e.g., Medical Record Number, Account Number, Unique Identifiable Number as determined by the facility, etc. A patient identifier is required for data submitted to the QIO Clinical Data Warehouse. ["Alphabetical Data Dictionary," versions 2.4, and 3.0 were effective during the audit period and defined the Data Element "hospital patient identifier" as stated, and the definition remained the same throughout the audit period.]

"Alphabetical Data Dictionary," version 3.0, page 1-291, indicates that the Data Element "Patient HIC#" is collected for CMS Only. It is collected by CMS for patients who have a standard HIC number. It is defined as the patient's Medicare health insurance claim number. Patient HIC# is required for data transmission of all cases that have a standard HIC#.


The Manual, version 3.0b, pages PN-2-1 and PN-2-2: Measure Information Form [Excerpt for Pneumonia]

Measure Set: Pneumonia (PN) Set Measurement ID #: PN-2 Performance Measure Name: Pneumococcal Vaccination

Description: Pneumonia patients, age 65 and older, who were screened for pneumococcal vaccine status and were administered the vaccine prior to discharge, if indicated.

Rationale: Pneumococcal vaccination is indicated for persons 65 years of age and older, because it is up to 75 percent effective in preventing pneumococcal bacteremia and meningitis. It is also an important vaccine due to increasing antibiotic resistance among pneumococci. In the United States today, vaccine coverage is suboptimal. Although inpatient vaccine screening and administration are recommended, hospitalization is an underutilized opportunity for adult vaccination.

Type of Measure: Process

Improvement Noted As: An increase in the rate

Numerator Statement: Patients with pneumonia, age 65 and older, who were screened for pneumococcal vaccine status and were vaccinated prior to discharge, if indicated

Included Populations: Not Applicable

Excluded Populations: None

Data Elements: Pneumococcal Vaccination Status

Denominator Statement: Pneumonia patients 65 years of age and older

Included Populations: Discharges:

• With an ICD-9-CM Principal Diagnosis Code of pneumonia as defined in [the Manual's] Appendix A, Table 3.1 ORICD-9-CM Principal Diagnosis Code of septicemia or respiratory failure (acute or chronic) as defined in [the Manual's] Appendix A, Tables 3.2 or 3.3

AND

• With an ICD-9-CM Other Diagnosis Code of pneumonia ([the Manual's] Appendix A, Table 3.1)

Excluded Populations:

• Patients less than 65 years of age • Patients who have a Length of Stay> 120 days • Patients with Cystic Fibrosis (Appendix A, Table 3.4)

Tallahassee Memorial HealthCare, Inc. , Quality Measure Data (A-04-1 1-08002) 7

• Patients who had no chest x-ray or CT scan that indicated abnormal findings within 24 hours prior to hospital arrival or anytime during this hospitalization

• Patients with Comfort Measures Only documented • Patients enrolled in clinical trials • Patients discharged/transferred to another hospital for inpatient care • Patients who left against medical advice or discontinued care • Patients who expired • Patients discharged/transferred to a Federal health care facility • Patients discharged/transferred to hospice

Data Elements:

• Admission Date • Birthdate • Chest X-ray • Clinical Trial • Comfort Measures Only • Discharge Date • Discharge Status • I CD-9-CM Other Diagnosis Codes • ICD-9-CM Principal Diagnosis Code

Data Collection Approach: Retrospective, data sources for required data elements include administrative data and medical record documents. Some hospitals may prefer to gather data concurrently by identifying patients in the population of interest. This approach provides opportunities for improvement at the point of care/service. However, complete documentation

includes the principal and other ICD-9-CM diagnosis and procedure codes, which require retrospective data entry.

Data Reported As: Aggregate rate generated from count data reported as a proportion.

Tallahassee Memorial HealthCare, Inc. , Quality Measure Data (A-04-1 1-08002) 8

APPENDIX B: PROCESS OF CARE QUALITY MEASURES

Heart Failure Process of Care Measures

Heart Failure (HF) is a weakening of the heart's pumping power. With heart failure, your body doesn't get enough oxygen and nutrients to meet its needs. These measures show some of the process of care provided for most adults with heart failure.

Set Measure %Met

ID# Measure Description Measure8

HF-1 Heart Failure Patients Given Discharge Instructions 90% Heart Failure Patients Given an Evaluation of Left Ventricular

HF-2 100%Systolic (L VS) Function Heart Failure Patients Given ACE Inhibitor or ARB for Left

HF-3 99%Ventricular Sy stolic Dysfunction (L VSD)

HF-4 Heart Failure Patients Given Smoking Cessation Advice/Counseling 100%

Heart Attack or Chest Pain Process of Care Measures

A heart attack (also called AMI or acute myocardial infarction) happens when the arteries leading to the heart become blocked and the blood supply is slowed or stopped. These measures show some of the process of care provided, if appropriate for most adults who have had a heart attack.

Set Measure %Met

ID# Measure Description Measure AMI-1 Heart Attack Patients Given Aspirin at Arrival 99% AMI-2 Heart Attack Patients Given Aspirin at Discharge 99%

Heart Attack Patients Given ACE Inhibitor or ARB for Left AMI-3 99%

Ventricular Systolic Dysfunction (L VSD) AMI-4 Heart Attack Patients Given Smoking Cessation Advice/Counseling 100% AMI-5 Heart Attack Patients Given Beta Blocker at Discharge 99%

Heart Attack Patients Given Fibrinolytic Medication Within 30 AMI-7a 0 patients*

Minutes of Arrival Heart Attack Patients Given Percutaneous Coronary Interventions

AMI-8a 90%(PCI) Within 90 Minutes Of Arrival

No Patients met the cnten a for mcluston m the measure calculatiOn.

8 As reflected on Hospita!Compare.hhs.gov as of December 11,2010.


http:Hospita!Compare.hhs.gov

Pneumonia Process of Care Measures

Pneumonia (PN) is a serious lung infection that causes difficulty breathing, fever, cough, and fatigue. These measures show some of the recommended treatments for pneumonia.

Set Measure %Met

ID# Measure Description Measure PN-2 Pneumonia Patients Assessed and Given Pneumonia Vaccination 98%

Pneumonia Patients Whose Initial Emergency Room Blood Culture PN-3 Was Performed Prior To The Administration Of The First Hospital 90%

Dose of Antibiotics PN-4 Pneumonia Patients Given Smoking Cessation Advice/Counseling 100%

Pneumonia Patients Given Initial Antibiotic(s) Within 6 Hours After PN-5 90%

Arrival PN-6 Pneumonia Patients Given the Most A12.propriate Initial Antibiotic(s) 91% PN-7 Pneumonia Patients Assessed and Given Influenza Vaccination 99%


APPENDIX C: AUDIT SCOPE AND METHODOLOGY

SCOPE

Our audit covered the Hospital's data submitted and used for Process of Care quality measures associated with HF, AMI, and PN in effect during the period July 1, 2009, through June 30, 2010. The specific quality measures reviewed are listed in Appendix B.

METHODOLOGY

To accomplish our objective, we:

• met with CMS program officials;

• reviewed applicable Federal laws and regulations and performance measure specification information;

• obtained quality measure case data from the Warehouse;

• obtained case discharge data from the Hospital;

• selected quality measures associated with HF, AMI, and PN to evaluate for completeness and accuracy;

• obtained HF, AMI, and PN measure information from HospitalCompare for the Hospital;

• interviewed Hospital staff and conducted an internal control walk -through of the steps involved in collecting, abstracting, and reporting quality measure data and in making corrections or adjustments to data;

• identified the Hospital's vendor and quality data abstraction software used for quality data submissions to the Warehouse and obtained a copy of the vendor contract;

• reviewed quarterly case trends for HF, AMI, and PN from the CMS Outcome Feedback Reports;

• tested the completeness of cases submitted to the Warehouse by selecting a judgmental sample of cases with HF, AMI, and PN diagnoses from Hospital cases and matching them to the Warehouse;

• tested the completeness ofthe Hospital cases by selecting a judgmental sample of Medicare cases with HF and AMI diagnoses from the National Claims History database and tracing them to the Hospital cases and to the Warehouse;

Tallahassee Memorial HealthCare, Inc. , Quality Measur e Data (A-04-11-08002) 11

• tested accuracy of CMS data by selecting a judgmental sample of cases with HF, AMI, and PN diagnoses from Hospital cases and comparing medical record details with Warehouse data abstraction details; and

• discussed the results of our review with Hospital officials.

We conducted this performance audit in accordance with generally accepted government auditing standards. Those standards require that we plan and perform the audit to obtain sufficient, appropriate evidence to provide a reasonable basis for our findings and conclusions based on our audit objectives. We believe that the evidence obtained provides a reasonable basis for our findings and conclusions based on our audit objectives.


,,... IJOO i\liccosul-l.'c Road

-''Tallaha~scc, f-l orida 32JO!i

850 43 1-1 155

Tallahassee Memorial Hospital Ttlllaha.Ut!e Me moria l 1/ea/tfi Care

July 8, 2013

Ms. Lori S. Pilcher Regional Inspector General Department of Health and Human Services (DHHS) Office of Inspector General (OIG) Office of Audit Services, Region IV 61 Forsyth Street, SW, Suite 3T41 Atlanta, GA 30303

Dear Ms. Pilcher:

On behalf of Tallahassee Memorial HealthCare, Inc. (TMH) and G. Mark O'Bryant,

President/CEO of TMH, please accept this letter in response to the findings made in Report

Number: A-04-11-08002 entitled Tallahassee Memorial HealthCare, Inc., Quality Measure Data

Were Not Always Complete and Accurate.

TMH constantly strives to develop and maintain systems which eliminate erroneous information

from its data collection processes. We note our understanding that the audit of TMH's quality

data submission was the first of its kind for the DHHS/OIG's Office of Audit Services. We think

this is sig nificant and contributed to the delayed completion of the audit as the OIG developed

it's understanding of the data collection and reporting processes as well as refinements

inevitably made to the audit work plan. We found the opportunity to participate in this important

review helpful to our goal of continuous improvement.

Finding: Incomplete Quality Measure Data Submission

The findings of the OIG audit state that data submissions made by TMH to the Centers for

Medicare and Medicaid Services (CMS) warehouse were incomplete as it relates to pneumonia

cases. Specifically the report stated that there were 14 cases in which the principal diagnosis

of septicemia or respiratory failure were present but that those cases were not in the CMS code warehouse data. TMH concurs with this finding.

Upon discovery of the 14 unidentified cases we conducted an investigation to determine the

source of the error. TMH expends considerable effort and resources to produce, review and

assemble the necessary quality metric data. The process includes review of many reports and data. In addition,individual patient charts in an attempt to accurately collect and report

recognizing the need for expert assistance in this data reporting TMH retains a nationally

recognized vendor that is certified by CMS to submit data on behalf of TMH. TMH reports its Reuters). quality data to CMS through Truven Health Analytics ("Truven") (f/k/a Thomson

Unfortunately, these 14 unreported cases were the result of an undetected software code error


ezygaText BoxAPPENDIX D: TALLAHASSEE MEMORIAL HEALTHCARE COMMENTS

in the database used by TMH to assemble the data. As a result, when the appropriate data was assembled from our patient records, the database software with the coding error incorrectly resequenced the order of the diagnosis codes from that contained in our source data. Complicating the reporting process, we also discovered that the software tool used by Truven to report data to CMS only accepts the first 24 diagnoses codes associated with a patient record. In each instance of the unreported 14 accounts, we determined the diagnosis codes for septicemia or respiratory failure were inadvertently re-sequenced which placed them lower than 24 other valid codes that were accepted and reported. Therefore, they were excluded from the dataset and thus not reported. We added steps to our data collection process to prevent this error from reoccurring. First, we corrected the software code error so the diagnosis codes are no longer re-sequenced prior to processing and second, we run additional queries for reportable cases based on diagnosis code to ensure the completeness of the data reported.

Although we concur with the finding of the 14 unidentified cases, we adamantly do not concur with the statement "The Hospital's management was unaware of its responsibilities for monitoring and reviewing data submissions for completeness; instead management relied upon the vendor to control the completeness and accuracy of data submitted to the Warehouse." We do not concur with this statement simply because it is not true; we engaged a vendor to provide additional expertise and assistance to a new, evolving reporting requirement with inherent complexities. We did not and do not rely solely upon our engagement with Truven to meet our reporting obligation. We have two full time staff who are devoted to assembling, reviewing and editing data prior to its submission to Truven for further transmission to the data Warehouse. In this instance, the data was reviewed and believed to be complete, we merely failed to identify and therefore did not report these 14 records out of 7,393 (0.19%) records due to an undetected software coding error in our database.

Report Finding: Inaccurate Quality Measure Data Submissions

The findings of the OIG audit state that TMH did not adhere to the specifications in terms of hospital patient identifier data. The report further states that the auditors could not trace back 100 percent of the cases submitted to the CMS data warehouse to the data provided by TMH. The report continues by stating that there were 24 heart failure and 25 acute myocardial infarction cases that did not reach the CMS data warehouse because the auditors could not confirm that the data was present in the data set provided to them, specifically the "Medicare Health Insurance Claim data field" contained "erroneous data as Health Maintenance Organization numbers and other non-Medicare insurance numbers". TMH does not concur with this finding.

The dataset reported to the Warehouse by TMH includes a unique CMS Tracking/Case 10 for 100% of the cases reported by TMH through Truven. The assignment of a unique 10 for each case is done to protect the privacy of the patient's information by removing other unique hospital identifiers. The CMS Tracking/Case 10 is assigned by Truven and a cross-walk table exists in the database to map the CMS/Tracking Case 10 to other hospital patient identifiers. Upon the initial production of data to the auditors, TMH inadvertently failed to provide the cross-walk which would have allowed the linking of each case to other identifiers. The cross-walk is available in the data reports and is readily available to TMH personnel. The cross-walk is not usually used by TMH personnel since they have access to the entire data set of all identifiers that, as noted, are matched to the CMS Tracking/Case 10.


2

With regard to the 24 heart failure cases and 25 acute myocardial infarction cases, upon investigation we subsequently determined the cases were correctly reported to the data Warehouse. The 49 cases at issue were admissions of patients who all belonged to a Medicare Advantage Plan (Part C). The Medicare Advantage Plan assigns a unique membership number for each member and that membership number was correctly used on the claim form submitted to the Medicare Advantage Plan HMO for the patient's care. The Medicare Advantage Plan HMO does not use the traditional Medicare Health Care Identifier. In addition, since TMH is a teaching hospital each admission of a Medicare Advantage Plan patient is also "shadow billed" to the CMS fiscal intermediary to notify CMS of the patient days of care to be used in determining TMH's graduate medical education payments. TMH did not report erroneous patient identifiers in the field referred to as "Medicare Health Insurance Claim". The Health Insurance Claim field contains the correct identifier which is the Medicare Advantage Plan HMO membership number in the instance of these 49 cases.

Finally, The CMS/T JC Specification Manual indicates that a Patient HIC number is required for data transmission of all cases that have a standard HIC. If patients do not have a standard HIC then this number cannot be sent to the CMS Warehouse. Truven's core measures software tool edits and flags for review any non-standard entries in the HIC field. Each non-standard entry in the HIC field is reviewed prior to submission to the data Warehouse and any necessary adjustments to correct or remove the HIC number are made prior to transmission. We believe the appropriate controls are in place to ensure all required data is completely and accurately submitted to the Warehouse.

Report Finding: Inadequate Policies and Procedures

The findings of the OIG audit specified TMH did not have adequate written policies and procedures to address and control the completeness and accuracy of the data submitted to the Warehouse. Although well established processes are in place, these are not formally documented. Therefore, TMH concurs with this finding.

TMH is in the process of creating a written policy and procedure to address the submission of data from the hospital to the Warehouse including the use of its vendor Truven Health Analytics. This policy and procedure will encompass the steps necessary to completely gather data from the various hospital information systems, verify the accuracy of the data gathered and insure transmission of the data is accomplished.

Thank you for the opportunity to comment on the findings and recommendations of this draft report. We hope that the additional information provided supports and will result in a reconsideration of those findings with which we do not concur. If you have any questions or need additional information, please let us know by contacting me at (850) 431-2136.

Respectfully,

Cynthia ~K~

L. Blair

Vice President/Chief Improvement and Planning Officer

Tallahassee Memorial HealthCare, Inc.


3

TABLE OF CONTENTSINTRODUCTIONWHY WE DID THIS REVIEWOBJECTIVEBACKGROUNDHospital Inpatient Quality Reporting ProgramPrior Work by the Government Accountability OfficeTallahassee Memorial HealthCare, Inc.HOW WE CONDUCTED THIS REVIEWFINDINGSINCOMPLETE QUALITY MEASURE DATA SUBMISSIONSINADEQUATE POLICIES AND PROCEDURESRECOMMENDATIONSTALLAHASSEE MEMORlAL HEALTHCARE COMMENTSOFFICE OF INSPECTOR GENERAL RESPONSEAPPENDIX A: FEDERAL REQUIREMENTSAPPENDIX B: PROCESS OF CARE QUALITY MEASURESAPPENDIX C: AUDIT SCOPE AND METHODOLOGY

executive summary why we did this review · executive summary tallahassee memorial healthcare,...

Documents